The "Information Age" is fundamentally altering the physical landscape of urban transportation. As commuters increasingly rely on real-time routing applications like Waze and Google Maps, and freight services move toward automation, traditional models of traffic flow are becoming obsolete. This paper, authored by a multi-institutional team from UC Berkeley and KTH Royal Institute of Technology, introduces a unified framework to model how information flow reshapes traffic patterns and proposes strategies to mitigate resulting inefficiencies.
A Multi-Layered Framework for Traffic Management
To address the complexities of modern mobility, the authors propose a hierarchical four-layer decision framework. This structure bridges the gap between long-term policy and real-time driver behavior.
- Social Planning: Focuses on long-term infrastructure and transportation policies, such as Stockholm’s congestion tax.
- Cooperative Control: Involves real-time management tools like dynamic tolls, priority lanes, and traffic light control.
- Routing Decisions: The "Information Layer" where routing applications suggest paths based on real-time data.
- Driver Behavior: The physical execution of travel, influenced by both personal decisions and external routing guidance.
Mathematical Modeling: Beyond Local PDEs
Traditional transportation engineering often treats traffic as a simple fluid using "local" Partial Differential Equations (PDEs), such as the Lighthill-Whitham-Richards (LWR) model. However, these models fail to account for varying levels of information among drivers.
The authors utilize non-local PDEs, where a vehicle's velocity is determined by the average traffic density over a stretch of road ahead. This approach more accurately reflects human driving behavior and prevents mathematical "shocks" or sudden density jumps. Furthermore, for freight coordination, the study employs distributed optimization and potential games. To maintain business privacy, homomorphic encryption (Paillier and ElGamal) is integrated, allowing competitive firms to find platooning opportunities without disclosing proprietary route data.
Case Study: The Nash Equilibrium on the I-210 Corridor
One of the most significant findings involves the "Waze Effect"—where selfish routing by individuals leads to system-wide congestion on local roads. Using simulation data from the I-210 corridor in Los Angeles, the researchers modeled the impact of routing apps on travel times.
As shown in the simulation, when approximately 17% of users adopt routing apps, travel times on the freeway and parallel arterial roads equalize. This represents a Nash Equilibrium: a state where no individual driver can reduce their travel time by switching routes. While this equilibrium may seem "fair," it is socially sub-optimal, as it diverts high volumes of traffic onto local streets not designed for such throughput.
Freight Optimization and Policy Impact
The framework was also applied to the Swedish road network to test freight platooning. The results were striking: by allowing for a mere 15-minute departure delay, the potential for fuel-saving platoons increased by 68%. A 30-minute delay window led to a 269% increase in platooning opportunities.
Additionally, the "Social Planning" layer was validated using a decade of historical data from Stockholm’s congestion charging system. The policy successfully reduced traffic by 20% initially and has sustained those levels despite urban growth, proving the efficacy of high-level policy interventions.
Conclusion
Information is no longer just a byproduct of transportation; it is a critical control variable. By understanding the interplay between information patterns and physical traffic, planners can design services that steer the system toward socially optimal outcomes. Future work will focus on reducing the computational complexity of non-local PDEs on large-scale city networks and refining models of human behavior to better account for the erratic nature of non-routed drivers.
Summary updated